AI agents are all the rage – and this time healthcare isn’t far behind other industries in adoption.
Healthcare traditionally trailed other industries in digital transformation and hospitals resist change, even with incentives. But this time, with big challenges ahead, health systems big and small are working on artificial intelligence and agentic AI solutions to revitalize the workplace, expand capacity, reduce stress, accelerate the revenue cycle and improve patient care. Rochester, Minn.-based Mayo Clinic has been on the forefront of AI for years and is leveraging agentic AI for a variety of functions.
“Building on the automation foundation in place across Mayo Clinic, we are now entering a bold new phase of innovation and impact,” said Anjali Bhagra, MD, medical director for automation at Mayo Clinic. “Our focus for the remainder of 2025 centers on pioneering and integrating agentic automation capabilities that seamlessly support both clinical and operational workflows.”
The next step for the health system will be developing an agentic automation architecture and integrated framework to scale and deploy intelligent agents. Dr. Bhagra and his team understand the importance of thoughtfully developing the technology for heightened security and sustainability. They’re also working on accelerating the next generation of human-AI interactions.
“By leveraging emerging technologies such as holographic interfaces and digital avatars, we are reimagining how care is delivered, experienced and supported: all anchored in our commitment to innovation, excellence and compassionate care,” said Dr. Bhagra.
Mount Sinai Health System in New York City is also creating an agentic AI roadmap focused on responsible, scalable AI aligning with strategic priorities. Robbie Freeman, chief digital transformation officer at Mount Sinai, told Becker’s the organization is co-designing solutions with frontline teams and building governance structures for safe and effective implementation.
“Our top focus for the second half of this year is advancing our digital strategy to better support our workforce, patients, consumers and care teams,” he said. “We’re prioritizing tools that enhance experience, streamline workflows and enable more personalized, coordinated care.”
AI agents are poised to transform the healthcare workplace, both for clinicians and operational teams. Darrell Keeling, PhD, senior vice president and chief information security officer at Fort Wayne, Ind.-based Parkview Health, said these tools could be part of the solution for workforce shortages and resource depletion; but they aren’t cheap.
“These tools come at a significant cost and must deliver value well beyond basic tasks like grammar correction, writing emails, creating presentations or creating a unique drawing to justify their return on investment,” said Dr. Keeling. “The real opportunity lies in training staff to thoughtfully integrate these agents into their daily workflows, not only to automate repetitive tasks but to enhance the value of their expertise.”
Many hospitals and health systems still have tight margins and are recovering – financially and operationally – from the COVID-19 pandemic. While it’s been shown that tech-forward hospitals adopting AI and meaningful digital transformation have bigger financial growth long-term, the upfront investment is still prohibitive.
“Can we cost-effectively invest in agentic agents without undermining our existing workforce? True success lies in leveraging the intellectual capital of our employees to guide, shape and optimize these technologies, not replace them,” said Dr. Keeling. “It’s more about reskilling our workforce and workforce augment.”
Lewis Marshall Jr., MD, chief medical officer at NYC Health + Hospitals / Lincoln Hospital is keeping a close eye on how AI and machine learning could quickly support the hospital with Medicaid cuts on the horizon. AI can automate the standard operating procedures, allowing the hospital to run efficiently with a lean team. AI can also help providers converse with patients, schedule appointments and identify specific diagnoses more accurately.
“We will need to look at agentic AI to make some decisions and take some actions, but we will need to make sure we have human oversight,” said Dr. Marshall. “While we are entering challenging times, there are potential solutions that, if identified, implemented and used thoughtfully, will reduce staff stress while improving patient satisfaction.”
Michael Laukaitis, director for revenue cycle analytics, accounting and quality assurance at UT Southwestern Medical Center in Texas, is also doubling down on agentic AI to evaluate inbound and outbound revenue cycle calls into an autonomous, voice-driven workflow. He’s using the technology for:
- Conversations with payers and patients
- Real-time voice engineering
- Zero-click documentation
“Every call is transcribed, summarized and written back to Epic and our CRM, giving staff instant context for the next touchpoint and shaving minutes off each interaction,” he said. “LLM-powered agents will verify eligibility, answer balance questions, start prior authorizations and even schedule follow-up calls while a human stays in the loop only when nuance is needed. We’re training custom acoustic models that understand regional accents and medical jargon, then pass intent and sentiment to our RPA bots so the next task fires before the caller hangs up.”
Integrating the AI agents effectively is easier said than done. AI literacy has become a big focus for many health systems, aiming to teach their teams about the technology and how to use it. Ideally, team members will hand over task-based responsibilities to AI agents and focus on the high-level, human-to-human interactions within their role.
“It’s not just about individual productivity,” said Dr. Keeling. “It’s about creating repeatable processes that others can adopt and build upon, enabling the enterprisewide adoption of agentic AI in a controlled and responsible manner.”
